Retatrutide's Phase 3 Results, Read Carefully: What Is Published, What Is Topline

Eli Lilly's retatrutide, an investigational injectable that activates GIP, GLP-1 and glucagon receptors, has now reported results from several Phase 3 trials, and the first full peer-reviewed paper is out. This explainer separates what has been published from what is still sponsor-reported topline, and lays out what the data do and do not show.

The short version

• Retatrutide is a single molecule that acts on three receptors (GIP, GLP-1 and glucagon). It is investigational and not approved anywhere. Lilly has said it plans to submit a Biologics License Application to FDA in Q1 2027.
• In TRIUMPH-1 (adults with obesity, no diabetes, 80 weeks), average weight change in the peer-reviewed NEJM paper was -17.6%, -23.7% and -25.0% across the three studied arms versus -3.9% for placebo. Lilly's topline efficacy estimand reported -19.0%, -25.9% and -28.3% for the same arms.
• TRIUMPH-2 (type 2 diabetes) and TRIUMPH-3 (severe obesity with established cardiovascular disease) met their primary endpoints. TRIUMPH-3 reported cardiovascular event counts, but the confidence intervals are wide and the data do not settle that question.
• Gastrointestinal events were the most common adverse events, and discontinuation due to adverse events rose in the higher arms.

Why this is getting attention

Retatrutide adds glucagon receptor activity to the GIP and GLP-1 activity found in dual agonists such as tirzepatide. Whether adding a third receptor target translates into better outcomes in people is what the TRIUMPH program is testing.

Lilly has described the program as more than 5,800 participants across the initial registrational trials, with further readouts expected over the next year. Four public announcements this year (May 21, June 6, July 23 and September 29) mean the numbers are spread across several press releases and one journal paper, which is why a single side-by-side view is useful.

What was studied

TRIUMPH-1 was a Phase 3, randomized, double-blind, placebo-controlled trial. Adults with obesity (or overweight with a weight-related condition) and without diabetes were assigned to retatrutide at one of three studied dose levels (low, middle or high) or placebo, for 80 weeks. The NEJM paper reports 2,339 randomized participants. Primary outcomes in the paper were percent change in body weight, change in WOMAC knee pain score in a subgroup of 574 participants with knee osteoarthritis, and change in apnea-hypopnea index in a subgroup of 243 participants with obstructive sleep apnea.

TRIUMPH-2 (NCT05929079) was a Phase 3, 80-week, randomized, double-blind, placebo-controlled trial in adults with type 2 diabetes and obesity or overweight, with 1,152 participants randomized 1:1:1:1 to one of three studied dose levels or placebo, according to the sponsor.

TRIUMPH-3 enrolled adults with severe obesity and established cardiovascular disease, with or without type 2 diabetes, comparing the middle and high dose levels with placebo over 80 weeks.

Key results

The table below puts the weight results side by side. The first row is from the peer-reviewed paper. The others are sponsor-reported topline figures, so treat them as preliminary until full publications are available.

Two details in that table matter when you read headlines. First, "28.3%" and "25.0%" describe the same arm of the same trial. The difference is the estimand. An efficacy estimand reflects outcomes as if participants stayed on treatment, while a treatment-regimen (intention-to-treat style) estimand includes everyone regardless of whether they stayed on drug. Neither is wrong, but they answer different questions, and the more conservative one is the better guide to what a real-world group might average.

Second, these are group averages. In the TRIUMPH-1 topline, Lilly reported that 45.3% of participants in the high-dose arm lost 30% or more of their body weight on the efficacy estimand, against 0.5% on placebo, and that 65.3% of the high-dose arm ended with a BMI under 30. A prespecified extension in participants with BMI of 35 or higher reported an average loss of 30.3% at 104 weeks in the high-dose group. That extension enrolled 532 participants who had completed the main study and tolerated their assigned dose, so it is a selected group.

Beyond weight, the NEJM paper reports improvements in knee osteoarthritis pain and sleep apnea severity. In the knee osteoarthritis subgroup, the WOMAC pain score changed by -3.2, -3.5 and -3.6 across the three arms versus -1.9 on placebo (hybrid estimand). In the sleep apnea subgroup, apnea-hypopnea index changed by -22.9, -34.3 and -31.7 events per hour versus -9.9 on placebo (hybrid estimand). Intention-to-treat figures are also reported in the paper.

In TRIUMPH-2, Lilly reported A1C reductions averaging 1.4% to 1.6% across arms from a baseline of 7.7%, versus 0.2% on placebo, with up to 40.0% of participants reaching an A1C below 5.7%. Lilly states that the TRIUMPH-2 results were presented at the 2026 EASD meeting and published simultaneously in The Lancet. We have linked the sponsor release below and have not independently reviewed the Lancet paper.

Safety and tolerability

Across the three trials, the most common adverse events were gastrointestinal, consistent in kind with other incretin-based therapies. In TRIUMPH-1, nausea occurred in 28.6%, 38.4% and 42.4% of the low, middle and high arms versus 14.8% on placebo, and vomiting in 10.6%, 22.8% and 25.3% versus 4.8%. Diarrhea and constipation followed a similar pattern.

Two events stand out because they are less familiar in this drug class. Dysesthesia (abnormal skin sensations) was reported in 5.1%, 12.3% and 12.5% of the TRIUMPH-1 arms versus 0.9% on placebo, and in 4.5% to 7.3% of retatrutide participants in TRIUMPH-2. Urinary tract infection was reported in 7.5% to 8.8% of retatrutide participants in TRIUMPH-1 versus 5.3% on placebo. The sponsor describes these events as generally mild to moderate, with most resolving during treatment. Detail on severity, time course and any follow-up will matter once full papers are available.

Discontinuation due to adverse events in TRIUMPH-1 was 4.1%, 6.9% and 11.3% across arms versus 4.9% on placebo. In TRIUMPH-2 it was 3.8%, 11.6% and 7.7% versus 4.9%, and in TRIUMPH-3 it was 9.8% and 13.5% versus 4.8%. In other words, the low arm in TRIUMPH-1 looked similar to placebo on this measure, and the higher arms did not.

What the data do not show

• Cardiovascular outcomes. In TRIUMPH-3, Lilly reported that major adverse cardiovascular events occurred less often than anticipated in both arms. For the five-component MACE endpoint the sponsor reported 44 events on retatrutide versus 52 on placebo (hazard ratio 0.82, 95% CI 0.55 to 1.22), and for the three-component endpoint 27 versus 23 events (hazard ratio 1.12, 95% CI 0.64 to 1.96). Both confidence intervals comfortably include 1.0. These counts are too few to show benefit or harm; the question would need a trial designed for it.
• Long-term durability. The main comparisons end at 80 weeks. The 104-week extension is a selected subgroup, and there are no data here on what happens after stopping or over many years.
• Real-world performance. Trial participants are screened, monitored and highly adherent compared with routine practice, and the trials ran under strict study protocols.
• Comparison with other agents. These were placebo-controlled studies. Cross-trial comparison with tirzepatide, semaglutide or oral agents is not valid, because populations, durations, estimands and baseline characteristics differ.
• Peer-reviewed detail for TRIUMPH-2 and TRIUMPH-3 beyond sponsor releases. Until those papers are read, subgroup and safety findings should be treated as topline.

Why it matters in clinic

For physicians and nurse practitioners, the practical value of this readout is in the framing, not a change in what is available. Retatrutide is investigational. It has not been approved, so anything marketed under its name today is not an FDA-approved product. FDA has separately warned about unapproved GLP-1 drugs, and the agency's published policy statements on compounding remain the relevant reference for questions about what can and cannot be supplied.

The data are most useful as context for patient conversations. Patients may have seen "triple agonist" headlines, and some may ask about it. A clinician who knows the difference between a topline press release and a peer-reviewed estimand, and who can say plainly that approval has not happened and the long-term picture is unknown, is better placed to answer. The same applies when a patient asks about a research peptide sold online that borrows the drug's name.

The size of the reported effect is also worth keeping in proportion. The peer-reviewed intention-to-treat figure (-25.0%) is lower than the headline topline figure (-28.3%), and tolerability, discontinuation and what happens after treatment stops are the areas where full papers and later studies will carry the evidence.

What to watch next

• Full publications for TRIUMPH-2 and TRIUMPH-3, including prespecified safety analyses and the characterization of dysesthesia.
• Lilly's planned FDA submission in Q1 2027, and what FDA says about the data package.
• Additional TRIUMPH readouts that Lilly has said are expected over the next year, including studies in other populations.
• Any dedicated cardiovascular or renal outcomes data, which is a separate question from weight loss.

How to read the sources

The list below labels each source by type. "Peer-reviewed" means a published journal article. "Sponsor-reported topline" means a company press release with figures that have not yet been through full peer review. Registry entries describe the trial design, not its results.

This article is an educational summary of published research and company-reported results. It is not medical advice, does not recommend any treatment, and is not an endorsement of any product.